We've seen too many high-profile calamities of websites going down in a major way—whether Target's, Ann Taylor's, Twitter's, or Facebook's. It's painful to watch, and certainly it must be far more painful to experience from the inside.
And, of course, such intermittent events aren't limited to retailers or big names, which are merely the ones that get the most press. Lost revenues and tarnished corporate and personal images don't discriminate by industry sector or company size. Online disasters can happen to anyone. And the last thing any of us wants or needs is bad press and irate customers.
More subtle than such blatant disasters, and in the long-term possibly more costly (and unnecessarily so), is poor or nonoptimal website performance that goes undetected.
Why Underperforming Websites Aren't Improved
By now, we know that degradation of website response times equates to lost revenue, and, similarly, improvements in website response times translate directly into revenue gains. For example, a two-second increase in response time reduced revenue per user 4.3% for Bing; and a five-second decrease in response time resulted in a 7-12% increase in revenue for Shopzilla.
You might have heard all this before... So why hasn't much of anything changed? Well, knowing of an issue is one thing, but having the time, authority, and tools to identify and fix the problem is something else entirely.
Many, if not most, marketers are understandably too busy with daily tasks and objectives (creating campaigns and tracking them, creating or increasing awareness, and maintaining or increasing inbound leads) to be able to properly focus on "customer perceived performance" (CPP). That lack of focus is particularly apparent among small and midsize businesses; but, even in large enterprises, too often the focus of Web analytics is on the more attention-getting aspects of it that don't deal with CPP.
As for the issue of authority, although CPP's impact is on Marketing, the determinant of it—Web performance—is generally housed within the IT department... Hello, efficient operation's old nemesis: departmental silos!
Why Marketing and IT Must Be Aligned
What is needed is a unified view. Marketing and IT should be looking at the same issue: that is, IT is constantly aware of Marketing's business considerations, and Marketing can see the real-time performance of IT operations that have an impact on both CPP and other Marketing concerns and initiatives. That cross-silo view can be achieved via the new generation of performance-visualization systems.
Having proper (i.e., engaged, enlightened, cooperative) alignment between the business side and IT is in line with the IT concept of business service management (BSM). However, coming to the level of BSM is a lengthy process with an uncertain outcome. BSM requires an IT department to, among other things, engage in significant personnel and process changes, including definition and documentation of services. Having a unified view is far speedier and more straightforward to implement and can even have a more profound and assured positive benefit.
Fortunately, from a prioritization of objectives point of view, aligning with business is one of the top priorities for CIOs, according to Gartner. And what better way for a CIO to facilitate such alignment than to have both IT and Marketing working off of the "same sheet of paper," or more correctly in this case, the same operational dashboards?
An additional benefit is that Marketing does not always have to be the bad guy, complaining to IT about this and that. IT will see, and be alerted to, the same issues that Marketing sees. IT can then start to share both your pain and your satisfaction.
Assessment of performance needs should not be limited to Web performance, the way Google Analytics is, for example. Why? Because you want IT to be able to quickly correlate the real-time information with its other systems, to do root-cause analysis. The days of switching back and forth between standalone systems that don't talk to one another are numbered. You want IT to be able to quickly work, with both historical and live graphs, to determine whether the CPP issue is caused by, say, Web servers, application servers, message queue systems, database servers, or caching servers.
When events or degradations like those described at the outset of this article happen, some element of the blame-game goes comes into play, yet no matter how much or how little Marketing bears the brunt of the blame, it is, of course, always incumbent upon us marketers to both have CPP recover quickly and work to make sure that such events are avoided in the future as much as possible. Hence the need for the root-cause analysis noted earlier.
So, in addition to rapid NOC (network operations center) response, as it's called in the IT/datacenter world, you want effective anticipation of emerging future needs and issues. Trending is one key way to do that. When both you and your IT department have graphs of your key operational metrics in front of you, you can see when things, say, take a downturn, become erratic, or increase. As for increases, they're not non-impact-producing; of course, even beneficial rises can result in greater than anticipated demand/draw on resources.
And so, no matter what type the anomaly being confronted, IT's being armed with this real-time knowledge can take pre-emptive action to ward off future hardship to you.
Why Aren't IT Departments Up to Speed?
"Why don't IT departments already use such a solution?" you might ask.
Well, for one thing, unified performance monitoring is a relatively new area and concept. Second, IT budgets have been cut to the bone, as with most budgets.
For the third reason, let's look at the typical causes of outages, say unplanned application downtime specifically. Estimates are that on average 40% are caused by operator errors, 40% by application failures (including "bugs" and changes to that cause unexpected problems), and only about 20% by hardware, operating system, environmental (heating, cooling, power, etc.), and disasters. Yet regarding monitoring, followed by application monitoring, that 20% seems to get the most attention. What's needed is a more holistic, comprehensive approach to performance monitoring that encompasses all areas, including Marketing.
And the fourth reason is that IT is not the one directly feeling the pain: Customers aren't complaining to IT, they're complaining to you, Marketing. So it's incumbent upon us as marketers to drive the implementation of cohesive performance monitoring.
So, as you can see, although monitoring has tended to be limited to the IT domain, that does not have to be the case. And there's plenty of reason for marketers to take advantage of new technologies to maximize customer perceived performance and minimize negative events on an ongoing basis.